Aniket Kulkarni

intrinsically curious

[Aniket Kulkarni @anikul94]

Getting started with AI agents: 10 concepts! A thread of threads ⬇️

[arnav @arnav_kumar]

Capitalism is indeed beautiful.
These are strategic moves of the highest order.
Why business at its highest level is 4D chess..

- How to corner markets?
- How to make the ecosystem lean in your favour?
- How to make sure no one in your ecosystem becomes powerful enough to dictate terms?
- How to prevent someone else from cornering adjacent markets that are valuable to you?

The strategy is "Commoditise your Compliment"

Simply put, if another company could capture excessive profits from a layer your business depends on, make that layer cheap or free—often by open-sourcing it

Some of the biggest examples:

Kubernetes – Google open-sourced Kubernetes. Google wanted to prevent AWS from becoming the proprietary operating system of the cloud.
By making container orchestration open and standard, Google helped commoditize an important layer above cloud infrastructure.

Microsoft - gave away Internet Explorer to commoditize browsers, protecting Windows.

Google - made Android free to commoditize mobile operating systems, protecting Search.

Open Compute Project (Facebook) – Facebook open-sourced its data center hardware designs. If server vendors earned excessive margins, Facebook would pay more. By creating an open ecosystem, it reduced infrastructure costs for everyone.

React (Facebook) – Facebook open-sourced React partly to ensure that the web remained a strong application platform rather than becoming dominated by proprietary mobile ecosystems.
A thriving React ecosystem also made it easier to hire engineers and build products.

Presto (Facebook) – Facebook open-sourced its distributed SQL query engine, creating a standard analytics platform instead of relying on expensive proprietary alternatives.

Absolute strategic drama ❤️

[Aniket Kulkarni @anikul94]

Who's winning the AI race? Well consumers for now :)

[Eidzoku @evi77ain]

Codex usage reset announcement in an hour or two🤞

[Aniket Kulkarni @anikul94]

@ClementDelangue 🫡

[clem 🤗 @ClementDelangue]

In the spirit of transparency, here’s what I asked @OpenAI:

• Radical transparency: let’s release the traces from the “rogue” agents so the entire research community can study what happened.

• More capabilities for defenders: let’s commit $100M in compute from OAI to help the Hugging Face community build powerful cyber defenses with the best open and closed models.

The first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!

[Aniket Kulkarni @anikul94]

I read something this week that I'm still trying to fully process.

@OpenAI was testing GPT-5.6 Sol and a more capable pre-release model on an advanced cyber security benchmark.

The models were supposed to solve the task inside a restricted environment.

Instead they found an unintended path out, reached the open internet, compromised parts of @huggingface's infrastructure while pursuing the benchmark objective.

Everyone's first reaction was obvious: "The model escaped the sandbox!"

But a more useful way to see this is this:

'The models stayed intensely focused on the goal. It just found a path humans had never intended to allow.'

That distinction matters.

The models were tested with some normal production cyber safeguards intentionally reduced, so OpenAI could measure their maximum cyber capability . Hugging Face later said that the activity was detected and contained, with no evidence that public models, datasets, spaces, packages, or container images had been tampered with.

This is not just a story about AI going 'rogue' or 'wild'. It is also a product-design problem.

We usually define the outcome we want from an agent, give it tools, and measure whether it completed the task. But capable agents may find routes that satisfy the metric while violating the intent.

That means success cannot only be about the final result.

It also has to include:

- Which actions are allowed?
- Where human approval is required?
- What the system can access?
- How its behavior is monitored?
- How quickly it can be interrupted ?

Because:

'The agent can complete the task and still fail the product.'

That is the part that stayed with me the most.

In normal software, success can often look binary: the workflow completed or it did not.

For an agentic product that is too weak. The outcomes matter but the path matters too.

There was one more interesting detail:

Hugging Face said, some hosted frontier models initially refused parts of the forensic analysis because the logs contained real exploit commands, payloads and credentials. The team eventually used an open-weight model (GLM 5.2) on its own infrastructure to continue the investigation, re construct the attack and help identify the root cause.

This suggests that safety systems also need better ways to distinguish legitimate defensive work from malicious use.

For builders the takeaway is:

Do not only define what the agent is allowed to do. Design for what happens when it finds a way around the boundary.

[Aniket Kulkarni @anikul94]

Calling "Open AI's agent observability broken" would be an overstatement, given how goated their team is!

They must have had the traces and telemetry in place.

The harder problem may be knowing which signals matter before the agent has already crossed the line.

https://t.co/9e6YiPHIQz

[Andrew Curran @AndrewCurran_]

New details about the Hugging Face incident from Reuters. The report says OpenAI noticed odd behavior before the event, including an agent leaving notes for future versions of itself with escape instructions.

[Aniket Kulkarni @anikul94]

@niketrajdwivedi So many 'leaks' jokes on my mind rn

[Aniket Kulkarni @anikul94]

@aayushchugh bro mogged vijay sales & croma, solid set up dude

[Aniket Kulkarni @anikul94]

@badbotvivi Another Steve jobs trait

[Aniket Kulkarni @anikul94]

💯

The latest @a16z pod has some insane @travisk gems which are helpful for anyone who is building anything!

The Semiconductor analogy that he gave for cloud kitchens was🔥

https://t.co/EF5Ijjp6wr

[anu @anuatluru]

Everyone is starved for a new Steve Jobs and Travis Kalanick is set up like he could be one. A rebel-visionary narrative, kicked out of his own company, back for the redemption/revenge arc with a big swing. A classic flavor of the greatest show on earth.

[Aniket Kulkarni @anikul94]

@amritwt GOATed team.

[Aniket Kulkarni @anikul94]

https://t.co/63RfNf9gMn

[Aniket Kulkarni @anikul94]

https://t.co/89dn7JRkFd

[Aniket Kulkarni @anikul94]

@Rahul_J_Mathur 🎯🎯🎯

[Aniket Kulkarni @anikul94]

@FarzaTV Cool stuff.

[Aniket Kulkarni @anikul94]

@prathamesss 💯

[Aniket Kulkarni @anikul94]

Let's go🚀 Let's see how it turns out for my project :)

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3: Open Frontier Intelligence

🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows

Kimi K3 is now live on on https://t.co/zrk6zZxZUo, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/YTfiMSNM1f

[@jason @Jason]

It’s happening folks… things accelerated more in the last 30 days from a dozen players than in the last year

Open Source models are compounding

Frontier Models are refining

When open source hits robotics, self driving and life sciences things are going to get wild

2026 will be the year it happened

WERE AT AGI IN 2026

SUPER INTELLIGENCE IS 2027-28

ITS GONNA GET VERY STRANGE

[Aniket Kulkarni @anikul94]

https://t.co/5bnxZ3Rd0z